CVE-2026-31951 LibreChat's MCP Server Header Injection Enables OAuth Token Theft
Attackers can steal OAuth tokens by creating malicious MCP servers with crafted headers, allowing account takeover in LibreChat.
- Exploited in the wild
- Not known
- Login needed
- Any signed-in user
- Attackable
- Over the internet
- Victim action needed
- A user has to click or open something
- Fix
- None listed
What is CVE-2026-31951?
CVE-2026-31951 is a medium-severity vulnerability in Librechat Librechat, published on 27 Mar 2026. Any signed-in user can exploit it over the network, and only if a user opens a file or follows a link.
Official description
LibreChat is a ChatGPT clone with additional features. In versions 0.8.2-rc1 through 0.8.3-rc1, user-created MCP (Model Context Protocol) servers can include arbitrary HTTP headers that undergo credential placeholder substitution. An attacker can create a malicious MCP server with headers containing `{{LIBRECHAT_OPENID_ACCESS_TOKEN}}` (and others), causing victims who call tools on that server to have their OAuth tokens exfiltrated. Version 0.8.3-rc2 fixes the issue.
Weakness: CWE-200
Which versions are affected by CVE-2026-31951?
It affects 1 product.
Products include Librechat Librechat.
How do I fix CVE-2026-31951?
No fix has been published for CVE-2026-31951 yet. Limit exposure of the affected software until one is available.
Is CVE-2026-31951 being exploited?
It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-31951 exists.
Other vulnerabilities in Librechat Librechat
- CVE-2026-32625 LibreChat Exfiltrates Server Secrets via MCP Server URL Injection
- CVE-2024-11167 Improper Access Control in danny-avila/librechat
- CVE-2026-22252 LibreChat MCP Stdio Remote Command Execution
- CVE-2025-69222 LibreChat is vulnerable to Server-Side Request Forgery due to missing restrictions
- CVE-2024-41704 Path traversal in n/a
Sources
Also known as GHSA-pmw7-gqwj-f954.